摘要
基于传统信赖域算法,为了解决约束最优化问题,利用非单调信赖域技术调整信赖域半径,矫正罚参数,提出了一种信赖域的修正算法,给出了收敛性证明.数值实验表明了算法的有效性.
Based on the traditional trust region algorithms,a kind of correction algorithm for constraint optimization problem was given.The algorithm adjusts the trust region radius by using nonmonotone trust region technology and corrects the penalty parameters,proves of the astringency were also given.At last,numerical experiment shows that the algorithm is efficient.
出处
《甘肃联合大学学报（自然科学版）》
2010年第5期16-20,共5页
Journal of Gansu Lianhe University :Natural Sciences
关键词
约束最优化
信赖域算法
非单调信赖域技术
罚参数
收敛性
constraint optimization
the trust region algorithm
nonmonotone trust region technology
penalty parameters
convergence
